Steam reforming of methanol over Cu/ZrO2/CeO2 …pubman.mpdl.mpg.de/pubman/item/escidoc:738831:1/component/escido… · Steam reforming of methanol over Cu/ZrO2/CeO2 catalysts: a

Journal of

J.Catal. 230, (2005), 464-475

Steam reforming of methanol over Cu/ZrO2/CeO2 catalysts: a kinetic study

A. Mastalir 1, 2, B. Frank 2, A. Szizybalski 1, H. Soerijanto 1, 2, A. Deshpande 3,

M. Niederberger 3, R. Schomäcker 2, R. Schlögl 1, T. Ressler 1, *

1Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG, Faradayweg 4-6, 14195 Berlin, Germany

2 Institut für Chemie, TU Berlin, Sekr. TC-8, Straße des 17. Juni 124, 10623 Berlin, Germany 3 Max-Planck-Institut für Kolloid- und Grenzflächenforschung, Abteilung Kolloidchemie,

MPI-KG 14424 Potsdam, Germany

* Corresponding author: e-mail [email protected], phone +49 30 8413 3192, fax +49 30 8413 4405 Abstract Steam reforming of methanol (SRM) was investigated over Cu/ZrO2/CeO2 (CZC) catalysts prepared via a novel synthetic method based on co-precipitation and polymer templating. Structural characterization of the samples was performed by N2 adsorption-desorption, N2O decomposition, and X-ray diffraction. The variation of the Cu loading resulted in an increased Cu crystallite size and a decreased specific surface area of the active particles. Catalytic investigations were carried out in a fixed bed reactor at 105 Pa, by applying a CH3OH:H2O = 1:1 ratio. The samples with Cu contents higher than 5 % exhibited good long-term stabilities and low CO levels during continuous operation. The kinetic model sug-gested for the transformation involved the reverse water-gas shift (RWGS) and methanol decomposition (MD), in addition to the SRM reaction. Kinetic measurements were accomplished in the temperature range 503-573 K and the experimental results could be well simulated. The highest methanol conversions and the lowest CO levels were observed in the temperature range 523-543 K. The apparent activation energies for the individual reactions were found to depend on the Cu content of the catalyst. Since the influence of mass transport limitations on the kinetic data could be excluded, it was established that the variation of the Cu concentration in the precursor material altered the microstructure of the Cu particles and, accordingly, the active Cu surface, which resulted in the formation of significantly different catalysts. Keywords: steam reforming of methanol, copper, zirconia, ceria, N2O chemisorption, long term stability, CO forma-tion, kinetic model, reverse water-gas shift reaction, methanol decomposition, activation energy 1.Introduction

In the past decade, considerable attention has been fo-cused on the reduction of the significant emissions originat-ing from mobile sources, such as internal combustion engines [1-4]. For environmental reasons, the development of proton-exchange membrane fuel cells (PEMFCs) has gained in increasing importance [5,6]. As compared with conventional heat engines, several advantages of fuel cell application have been established, including a higher effi-ciency and a more convenient operation, the absence of moving parts and the low emission of hazardous compounds [1,5]. The combustion of hydrogen in a fuel cell is regarded as a clean process, releasing energy and providing only wa-

ter as an exhaust material [4,7,8]. However, hydrogen is not a natural energy source and must be generated by consuming a large amount of energy, either from natural gas or via the electrolysis of water [4]. Furthermore, for a fuel cell vehicle, the storage and the supply of hydrogen, a volatile and explo-sive gas, imposes mechanical problems and safety hazards on a commercial level [1,4,8].

Several liquid fuel candidates have been discussed for on-board reforming, including methanol, ethanol, gasoline and diesel [1], of which methanol is considered the most favourable alternative [1,9]. Although mostly produced from natural gas, methanol can also be obtained from renewable sources, thus lowering the production of carbon dioxide [5,9]. Moreover, via methanol synthesis from hydrogen and carbon dioxide, methanol can be employed as a storage and

Preprint of the Department of Inorganic Chemistry, Fritz-Haber-Institute of the MPG (for personal use only) (www.fhi-berlin.mpg.de/ac)

Steam reforming of methanol over Cu/ZrO2/CeO2 catalysts: a kinetic study, A. Mastalir et al., J.Catal. 230, (2005), 464-475


2

transport medium for hydrogen. In addition to its good availability and low boiling point, methanol possesses a high hydrogen to carbon ratio (4:1) and contains no carbon-carbon bond, which considerably reduces the risk of coke formation during its reactions [1,5,6].

The most important processes to produce hydrogen from methanol (Eqs.1-3) are decomposition (MD), steam reforming (SRM), and partial oxidation (POM) [1]. The combined or oxidative reforming of methanol (CRM or OSRM), a combination of (2) and (3), has also been investi-gated [2,6,10,11]. The decomposition of methanol is a strongly endothermic reaction producing a high CO yield, which makes it rather unsuitable for fuel cell applications [1]. The steam reforming reaction is also endothermic, how-ever, it typically affords a substantial H2 yield of 75%, while maintaining high carbon dioxide selectivity (∼ 25%) [1,4,8]. The partial oxidation of methanol is an exothermic reaction with a rapid start-up and a dynamic response [4]. Neverthe-less, the formation of hot spots in the reactor may result in sintering of the Cu particles, which tends to decrease the catalytic performance [8,11]. Furthermore, this reaction produces a considerably lower amount of hydrogen than SRM [1,8,9].

As the highest hydrogen yield may be achieved by the steam reforming of methanol [1], the latter reaction was investigated within the framework of the present study. The major drawback of SRM is the formation of CO as a by-product, which, even at a low concentration of 100 ppm, decreases the fuel cell performance by poisoning the Pt elec-trode [2,6,12]. Currently, second stage catalytic reactors are being used to remove CO by the water-gas shift reaction, oxidation or methane formation [6,10]. However, this CO clean-up unit is rather inconvenient as it occupies a large volume and decreases the efficiency of the fuel cell through hydrogen consumption [8]. In order to eliminate the need for gas purification, high performance catalysts are required that provide substantial methanol conversion and H2 selectivity, together with the lowest possible amount of CO [10,12]. The complete absence of CO in the product gas may be difficult to achieve under SRM conditions.

Several studies have been reported in the literature on the applications of Cu- and Pd-based catalysts for SRM [1-6,8-12,13-18], of which Cu containing catalysts are clearly preferred because of their high activity and selectivity at lower temperatures [1,12,19]. Although Cu catalysts are also regarded as susceptible to thermal deactivation, their sinter-ing abilities may be considerably reduced by the addition of one or more oxide species, such as ZnO, Al2O3 or Cr2O3 [20]. The most efficient catalysts for SRM, including the traditional Cu/ZnO and Cu/ZnO/Al2O3 systems, have been

investigated in detail [1,3,6,9,15,16,21-25]. The in situ char-acterization of Cu/ZnO under SRM activation and operating conditions revealed that the interaction of the Cu and ZnO phases has a pronounced effect on the catalytic activity [22,23]. For Cu/ZnO/Al2O3 catalysts, high methanol conver-sions and low CO selectivities have been obtained [1,5,8,9,16] and kinetic analysis provided important infor-mation on the overall reaction mechanism [5,6,16,21,25,26]. Compared to the conventional ZnO or Al2O3-supported Cu catalysts, ZrO2-containing samples have shown increased activities and reduced CO levels for the SRM reaction [3,5,12,27-29]. The promoting effect of ZrO2 has been at-tributed to an improved reducibility of CuO, which tends to increase the Cu dispersion [29]. ZrO2 has also been reported to prevent the sintering of Cu crystallites under reaction conditions [5,12,29] and thus may be regarded as a structural stabilizer. Likewise, the application of CeO2, as either a support material or a promoter, has been found to improve the efficiency of Cu-based catalysts [11,30,31]. CeO2 has been found to increase the thermal stability and the activity of Al2O3-supported Cu catalysts through a synergetic effect and favour the conversion of CO via the water-gas shift reaction [30]. Moreover, CeO2 has a high oxygen storage capacity [32,33] and the partially reduced CeO2 sites formed under reductive conditions produce mobile oxygen, which tends to have a beneficial effect on the catalytic performance [31,32]. For ZrO2-CeO2 supported catalysts, the interaction between the oxide phases, leading to the formation of a thermally stable solid solution, has been reported to increase the mobility of oxygen in both phases [32].

D

D

R

R

O

O

k0 1

3 2k

k0 1

3 2 2 2k

k0 1

3 2 2 2k

CH OH(g) 2H CO H 91 kJmol (1)

CH OH(g) H O(g) CO 3H H 50 kJmol (2)

CH OH(g) 0.5O 2H CO H 192 kJmol (3)

−

−

−

−

−

−

+ ∆ =

+ + ∆ =

+ + ∆ = −

The aim of the current work was the structural and catalytic investigation of novel Cu/ ZrO2/CeO2 catalysts prepared by co-precipitation. In order to increase the specific surface area, highly porous polymer beads were employed as template for the preparation of the solid catalysts [34]. Sam-ples with different Cu loadings were examined for the steam reforming of methanol and their catalytic properties are discussed, including long-term stability and CO production. Furthermore, the reaction mechanism was studied and a kinetic analysis was undertaken to determine the rate con-stants and the activation energies for the model reactions. 2.Experimental 2.1.Catalyst preparation

The Cu/ZrO2/CeO2 (CZC) catalysts were synthesized from metal sols, prepared from the appropriate amounts of the mixed precursors (NH4)2Ce(NO3)6, ZrO(NO3)2⋅H2O and Cu(NO3)2⋅2.5 H2O (all Aldrich products of > 99% purity), dissolved in 40 ml of distilled water. The total metal content of the resulting sol was 5⋅10-3 mol. Co-precipitation of the metal sols as metal hydroxides was accomplished by in-creasing the pH above 10 through addition of a NaOH solu-tion. The resulting precipitate was repeatedly washed with distilled water until pH=7 was obtained, and then suspended in a mixture of 5 ml of distilled water and a calculated



3

amount of 90% HNO3. Depending on the sample composi-tion, the HNO3:metal molar ratio was varied between 1:1 and 1.5:1. Ultrasonic treatment of the suspension for 45-60 min resulted in the formation of a transparent sol [35].

Templating was carried out by using XAD-16 non-functionalized polystyrene beads (a Sigma product), with a specific surface area of 800 m2g-1 and an average pore di-ameter of 10 nm. 1.5 ml of the wet beads and 0.005 mol of the metal sol were mixed together and then heated in an oven at 333 K for 2 days in order to ensure solvent evapora-tion. The beads were subsequently dried at 373 K for an-other day and then washed with distilled water to remove any excess deposit from their surface. After drying at 333 K for 1 day, the samples were calcined in an oven at 773 K in a N2 stream of 100 ml/min for 2.5 h and then for an additional 8 h in air. Upon calcination grey-green materials were ob-tained (see Fig.1). Due to the mean particle diameter (500 µm) and the regular character of the polymer beads, the resulting CZC samples could be conveniently applied as catalysts in fixed bed reactors.

Figure 1: Scanning electron micrograph of a Cu/ZrO2/CeO2 catalyst (CZC5) after calcination.

Four representative CZC samples were synthesized by the above method, with Cu contents 5, 15, 25 and 35%, de-noted as CZC5, CZC15, CZC25 and CZC35, respectively, to be characterized and tested for the steam reforming of methanol. The molar ratio ZrO2:CeO2 = 1:1 for the support materials was maintained in each case. 2.2.Characterization 2.2.1. X-ray fluorescence analysis (XFA)

The results of the elementary analysis of the CZC samples, performed by a Seiko (SEA 2010) instrument, are

shown in Table 1. The amount of Cu detected in the samples was found to be only slightly lower than that expected from the preparation conditions and the ratio ZrO2:CeO2 = 1:1 proved to be well adjusted. 2.2.2. N2 adsorption-desorption

The specific surface areas of the CZC samples were determined from N2 adsorption-desorption data obtained at the temperature of liquid N2 (77K), by using a Micromeritics 2375 BET apparatus equipped with a Vacprep 061 degasser. In order to remove traces of water and impurities from the surface, all samples were degassed before measurement at 13 Pa and 393 K for 12 h. The specific surface areas were calculated from the BET equation and the average pore di-ameters dp were obtained by the BJH method from the de-sorption branches of the adsorption isotherms. 2.2.3. N2O decomposition

The specific Cu0 surface areas of the samples were measured by N2O decomposition, by using the reactive fron-tal chromatography (RFC) method introduced by Chinchen et al [36].

The method is based on the oxidation of the exposed Cu0 surface atoms of a reduced sample by switching the feed from an inert gas stream to a mixture containing N2O as an oxidizing species. The Cu0 surface area can then be deter-mined from the amount of N2O consumed in the reaction, as given below (Eq.4).

)4(""2 222 NOCuONCu +→+

Before measurements, the CZC samples were acti-vated in a stream of MeOH:H2O = 2:1, with a flow rate of 36 cm3min-1 at 523 K. In order to diminish the amount of adsorbed molecules on the Cu surface, the samples were subsequently purged in pure He (50 cm3min-1) at 523 K for 1 h and then cooled down to 313 K, where the measure-ments were performed in a 0.5% N2O/He stream of 15 cm3min-1. The catalysts were diluted with boron nitride to provide a sample bed of ca. 25 mm height, with a thermo-couple positioned directly inside the powder bed. The mate-rial was placed onto a quartz frit in a tubular quartz reactor.

For the RFC method described by Chinchen et al., the area under the ion current trace (m/e = 28 for N2) has been directly used to determine the Cu0 surface area. For the cur-rent experiments, however, the same MS signal obtained from the amount of N2 produced for the CZC samples was found to be insufficient for evaluation. Therefore, the amount of N2 formed during decomposition was indirectly estimated from the retarded evolution of N2O (m/e = 44), as compared with blank experiments performed with boron nitride. Earlier investigations for Cu/ZnO samples revealed a good agreement between the Cu0 surface areas determined by N2O RFC with those obtained from the Cu crystallite size [23,37]. The Cu0 surface area was calculated by assuming a



4

stoichiometry of N2O/Cu = 0.5 and a value of 1.47⋅1019 Cu atoms/m2 for the surface density [5,29,36]. 2.2.4. X-ray diffraction (XRD)

Ex situ XRD measurements were performed on a STOE STADI P transmission diffractometer (CuKα radia-tion, Ge monochromator), equipped with a position sensitive detector (PSD, internal resolution 2Θ = 0.01°). The CuO crystallite diameters were calculated from the Scherrer equa-tion [38]. The full width at half maximum (FWHM) was determined by fitting a Lorentzian profile function to the CuO (111) and (200) diffraction peaks at 2Θ = 38.371 and 38.923°, respectively. 2.3. Catalytic test reaction

The steam reforming of methanol was studied in a fixed bed tubular reactor (stainless steel, id = 10 mm), by utilizing a 3-channel set-up, which ensured that 3 catalysts could be investigated at the same time, under exactly the same conditions. In order to achieve an efficient heat trans-fer, the reactors were placed in an aluminium heating block. The temperature of the reactor was regulated by PID control of the cartridge heaters situated inside the aluminium block. Time-on-stream investigations were performed at 523 K and the mass of catalyst was varied between 0.100 and 0.770 g, depending on the Cu content, in order to keep the amount of CuO constant for all samples (0.0225 g). For kinetic meas-urements, the mass of catalyst was increased by a factor of 2.6, corresponding to a CuO content of 0.0584 g for each sample, to achieve full conversion in the temperature range investigated (503-573 K). The catalyst was supported inside the reactor by a stainless steel grid. For flow conditioning, inert Pyrex beads (d = 500 µm) were placed below and on top of the catalyst bed. The reactant mixture of MeOH (HPLC grade, 99.9% purity) and distilled water, with a mo-lar ratio MeOH:H2O = 1:1, was introduced into the reactor by means of a Dionex HPLC P 580 pump. Both reactants

were degassed before use at 20 kPa. For time-on-stream measurements, the flow rate for a single channel was 0.07 cm3min-1, whereas for kinetic investigations, it was varied between 3.33⋅10-3 and 0.67 cm3min-1. Kinetic measurements were performed by using fresh catalysts to eliminate aging effects and experimental data were collected after 5 days time-on-stream. In order to achieve steady-state conditions, experimental data at low flow rates were collected 60 min after the flow rate had been adjusted. Before measurement, the catalysts were activated in situ in a MeOH:H2O = 1:1 stream of 0.07 cm3min-1 at 523 K for 16 h [15,16,25].

The reaction products consisting of effluent gases (H2 and CO2 as main products and a minor amount of CO), to-gether with unreacted methanol and water, were passed through a condenser at 268 K, which removed most of the liquid from the product gas stream. Complete removal of the condensed reactant mixture was achieved by applying an additional condenser at 253 K. The liquid composition was analyzed by using an Intersmat gas chromatograph (GC), equipped with a thermal conductivity detector (TCD), oper-ated with a 50m x 0.53 mm CP-Wax column at 363 K. The composition of the dry product gas mixture was determined on-line, by means of a Varian 3800 GC with a TCD. Helium was applied as a carrier gas and separation was achieved by using a 25 m x 0.53 mm CarboPLOT P7 column at 304 K. For quantitative analysis of the product gas stream, a cali-bration gas mixture, containing 0.5% CO, 4.5% N2, 25% CO2 and 70% H2 (a Messer Griesheim product), was em-ployed. Under the present experimental conditions, no evi-dence for the formation of the by-products methane, methyl formate and dimethyl ether was found [27,39]. 3.Results and Discussion 3.1.Catalyst characterization and long-term stability

The results obtained for the CZC samples from N2 ad-sorption-desorption and N2O decomposition are summarized in Table 1.

Table 1: Characteristic data of the CZC catalysts

Sample Cu

[%]

Zr

[%]

Ce

[%]

BET area

[m2g-1]

dp

[nm]

Cu0 area

[m2g-1catalyst]

Cu0 area

[m2g-1Cu]

CZC5 4.4 44.6 51.0 96 9.4 0.9 51.4

CZC15 12.1 44.3 43.6 102 9.9 1.8 25.8

CZC25 23.9 37.6 38.5 94 9.8 1.8 12.2

CZC35 31.0 32.7 36.3 83 10.0 1.5 7.6



5

The BET surface areas obtained for the CZC samples were higher than those reported for ZnO supported Cu cata-lysts [39] and of a similar order as those published for Cu/Al2O3 samples by Cheng et al. [19]. A decreasing BET surface area with increasing Cu loadings [19] was observed only for the samples with Cu contents exceeding 5%. As seen in Table 1, the values for the average pore diameter were very close for each sample (9.35-10.0 nm). Although the Cu0 specific surface areas for the CZC materials proved to be lower than expected from the internal surface areas and the Cu contents [29], the values exhibited a similar trend as the BET results. The highest Cu0 surface area per gram copper was obtained for CZC5. An increasing Cu content resulted in a decreasing Cu0 surface area roughly in agreement with an increasing Cu particle size. The net de-crease of the Cu0 specific surface area with increasing Cu content, observed for CZC15, CZC25 and CZC35, is in correlation with earlier results reported in the literature for supported Cu catalysts [5,19].

CZC5CZC15CZC25CZC35

0

1000

2000

3000

4000

10 30 50 70 90 110

2Θ [°]

Cou

nts

Figure 2: X-ray diffraction patterns of four CZC catalysts with a nominal copper concentration of 5 % (CZC5), 15 % (CZC15), 25 % (CZC25), and 35 % (CZC35) with a con-stant ZrO2/CeO2 ratio of about 1. Arrows indicate CuO phase.

The XRD patterns of the CZC samples displayed in

Fig.2 revealed that the characteristic CuO peaks could be distinguished only for the samples with the highest Cu con-tents. The absence of CuO signals for CZC5 and CZC15 may be attributed in part to their lower Cu loadings. Fur-thermore, the Cu crystallite size for the above samples may be too small to be detected by XRD, as related to a pro-nounced line broadening [5]. The broadening of the peaks also renders it difficult to establish whether the Zr, Ce and O atoms are arranged in one single mixed phase or in sepa-rate phases. For the same reason, the corresponding crystal structures (cubic and/or tetragonal) cannot be reliably de-termined. The Cu crystallite sizes obtained from the dif-fraction peaks for CZC25 and CZC35 were 12.4 and 15.3 nm, respectively. The slight increase in the particle size observed for the latter sample is in line with the decrease of the Cu0 specific surface area obtained from N2O decompo-sition. Further structural data of the Cu/CeO2/ZrO2 catalysts

obtained under reaction conditions and corresponding structure-activity relationships will be presented elsewhere.

In order to make a comparison between the long-term stabilities of the CZC catalysts, the methanol conver-sions obtained during a continuous operation of 16 days under standard conditions were plotted as a function of time-on-stream (Fig.3).

0

0.1

0.2

0.3

0.4

0 2 4 6 8 10 12 14 16 18

time-on-stream [day]

Con

vers

ion CZC5

CZC15CZC25CZC35

Figure 3: Methanol conversions during SRM of four CZC catalysts as a function of time-on-stream (mCuO = 0.0225 g, T = 523 K, p = 105 Pa, w = 0.07 cm3min-1, MeOH:H2O = 1:1)

It can be seen that no linear correlation exists be-

tween the catalytic activity depicted in Fig. 3 and the Cu surface area (Table 1). Evidently, the catalytic activity of the CZC samples is not only determined by the the Cu sur-face area, but also by the particular microstructure of the Cu particles, which may be similar to the microstrain in Cu/ZnO catalysts [23]. The catalytic activities of all the samples were found to diminish considerably for a period of 5 days, although to a different extent, which displayed no systematic variation with the Cu content. The conver-sions determined after 5 days approached a constant value, except for CZC5, which exhibited a further decrease. After an initial period of deactivation, the CZC samples with Cu contents exceeding 5% displayed good long-term stabilities and, hence, these samples may be suitable for extended operation. The relatively poor thermal stability of CZC5 may be related to the small Cu particle size, considering that small particles tend to be more susceptible to thermal sintering under reaction conditions [40], even at tempera-tures lower than 523 K [41]. In addition to the CZC sam-ples, a copper-free CeO2/ZrO2 sample was prepared by using the same procedure as described above. This material did not exhibit any detectable activity in the steam reform-ing of methanol under the reaction conditions employed. Hence, the differences observed for the CZC catalysts in the steam reforming of methanol can be mainly attributed to the bulk and surface properties of the Cu phase in these materials.



6

The CO partial pressures determined during the above measurements are depicted in Fig. 4. As mentioned above, the sample masses were adjusted to provide the same Cu content. For this reason, the mass of CZC5 con-siderably exceeded those of the other samples, and this may accounted for the amount of CO detected for CZC5 being substantially higher than those for the samples with higher Cu contents. Moreover, following a pronounced decrease for about 5 days, the CO levels were stabilized and did not exhibit any significant difference. This finding is consistent with the results of N2O decomposition, indicating that the active Cu contents of these samples were very similar. It may be concluded from Figs. 3 and 4 that an increase of the Cu content from 5 to 15% has a beneficial effect on the catalytic performance of CZC through improving the long-term stability and suppressing the CO production. Although the effect of a further increase of the Cu loading proved to be less important for the stability of the catalyst, it resulted in lower CO levels. Nevertheless, the CO production is also dependent on the conversion of methanol [29], as the activ-ity order of the CZC samples (Fig. 3) is in correlation with the order of the CO levels throughout the entire time-on-stream interval (Fig.4). Accordingly, under the present experimental conditions, the samples with increased Cu contents may be regarded as more efficient catalysts for SRM, with respect to the formation of CO in particular.

0

40

80

120

160

0 2 4 6 8 10 12 14 16 18

time-on-stream [day]

p CO [P

a] CZC5CZC15CZC25CZC35

Figure 4: CO formation during SRM of four CZC catalysts as a function of time-on-stream (mCuO = 0.0225 g, T = 523 K, p = 105 Pa, w = 0.07 cm3min-1, MeOH:H2O = 1:1)

Considering that the CZC catalysts are porous mate-rials and the methanol conversions in Fig.3 displayed no systematic variation with either the Cu loading or the Cu0 surface area, the effect of mass transport limitations on the catalytic performance was additionally investigated. It was previously observed that the CZC samples were not uni-form in size. Thus, for each material, three fractions with different particle diameters could be obtained by sieving. Moreover, a Cu/CeO2 sample containing 25% of Cu, pre-pared by the same method, was found to slightly deteriorate during SRM, as a small amount of the original catalyst beads (3.3%) was transformed into a fine powder after

reaction. On the other hand, the Zr-containing beads of the CZC samples remained essentially unchanged after cata-lytic investigations, which confirms that ZrO2 is an impor-tant structural stabilizer [12,29]. The methanol conversions obtained for the different sieve fractions of CZC15 under the same conditions are plotted in Fig.5 as a function of the reactant flow rate.

0.3

0.45

0.6

0.75

0.9

0 0.1 0.2 0.3

flow rate [cm3min-1]C

onve

rsio

n

100 - 300 µm

450 - 500 µm

650 - 710 µm

Figure 5: Methanol conversions during SRM as a function of the reactant flow rate, determined for various sieve frac-tions of CZC15 (m = 0.5500 g, T = 523 K, p = 105 Pa, MeOH:H2O = 1:1)

On increasing the flow rate, the conversions for the

various sieve fractions declined to the same extent. At a given flow rate, the values obtained slightly decreased with an increasing particle diameter, however, the differences detected are not significant with respect to mass transport limitation. This is corroborated by the dependence of the CO partial pressure on the methanol conversion, as de-picted in Fig.6.

0

100

200

300

400

500

600

0.25 0.5 0.75 1

Conversion

p CO [P

a] 100 - 300 µm

450 - 500 µm

650 - 710 µm

Figure 6: CO partial pressures determined for various sieve fractions of CZC15 during SRM under standard con-ditions (m = 0.5500 g, T = 523 K, p = 105 Pa, MeOH:H2O = 1:1)



7

According to Fig.6, the CO levels for the three frac-tions at the same conversions were very similar and thus all values could be fitted with the same exponential function. It may therefore be concluded that intraparticle transport phe-nomena exerted no considerable influence on the CO for-mation. Hence, because mass transport limitations can be excluded, a reliable kinetic model was suggested for the steam reforming of methanol. The kinetic parameters were determined by fitting the model to the experimental results obtained for the CZC catalysts. 3.2. Kinetic model

The individual reactions to be included in the kinetic model of the methanol steam reforming process are still under debate. Earlier studies have suggested that the kinet-ics could be sufficiently described by using only one or two of the possible overall reactions while assuming that the others were either at equilibrium or their rates were negli-gible. Initially, SRM was supposed to proceed by the for-mation of CO and H2, followed by the water-gas shift (WGS) reaction [15,42]. The formation of CO2 by the di-rect reaction of methanol and steam has also been proposed [43]. Methanol decomposition was involved in the reaction mechanism in certain studies [14,42,44], whereas in other cases, it was regarded as insignificant [16,21].

An adequate kinetic model for the steam reforming of methanol on a commercial Cu/ZnO/Al2O3 catalyst, which includes methanol decomposition (Eq.1) and the WGS reaction (Eq.5), in addition to the SRM process (Eq.2), has been recently suggested by Peppley et al [25].

W

W

k0 1

2 2 2k

CO H O(g) CO H H 41 kJmol (5)−

−+ + ∆ = −

Further investigations on SRM revealed that the

above model may also apply for other catalysts including Zr- or Ce-containing systems [28-30]. For the present study, it was found that a kinetic evaluation performed by including MD in the reaction scheme afforded a better fit to the experimental data. Accordingly, evaluation of the ex-perimental data obtained for the CZC samples was based on Peppley’s model. The kinetic model suggested by Taka-hashi et al., involving a methyl formate intermediate [45], may be regarded as less suitable for evaluation, given that no methyl formate production was observed under the pre-sent experimental conditions. In fact, methyl formate is more likely to be formed when the reaction is conducted in an excess of methanol [3].

CO formation has been generally observed at high methanol conversions and long contact times, indicating that CO is a secondary product, formed by the reverse wa-ter-gas shift (RWGS) reaction [3,6,28]. For the CZC sam-ples, kinetic analysis suggested that the amount of CO produced by MD was considerably lower than that formed by the RWGS reaction and therefore the reaction scheme was simplified by regarding MD as an irreversible reaction.

Furthermore, because of the low Hüttig temperature of Cu [46], as reflected in its relatively low melting point (1356 K) [5,6], Cu-based catalysts tend to undergo thermal deac-tivation at temperatures exceeding 573-623 K [5,6,47]. In order to eliminate ageing effects, kinetic measurements for the CZC samples were carried out in the temperature range 503-573 K.

According to the above considerations, the reaction scheme utilized for methanol steam reforming on CZC catalysts included the MD, SRM and the RWGS reactions (Eqs.1,2,5). Because SRM is the sum of the other two reac-tions, the above processes are not independent [26]. The partial pressures of the reactants and products were calcu-lated by assuming that the contribution of MD to the con-version was negligible. If the SRM reaction for an equimolar mixture of methanol and water takes place with a conversion of X, then the total mole number after SRM is n = n0(1+X). By considering that the RWGS reaction pro-ceeds with a conversion of Xw, the partial pressures of the reactants and products can be expressed as follows:

2

2

2

wCO

wCO

wH

MeOH

wH O

X Xp p (2 2XX (1 X )

6)

p p (2 2X

X (2 X )

7)

p p (2 2X1 X

8)

p p (2 2X

1 (1 X )X

9)

p p (10)2 2X

⋅= ⋅

+⋅ −

= ⋅+

⋅ += ⋅

+−

= ⋅+

− −= ⋅

+

The partial pressures can be obtained from Eqs. 6-10, given that p = 105 Pa, and Xw can be calculated from Raoult’s law as

2

COw

CO CO

X (11)ϕ=

ϕ + ϕ

where COϕ and 2COϕ are the volume fractions for CO and

CO2, respectively, which can be determined from the corre-sponding GC peak areas after analyzing the calibration gas mixture.

The rate equations for the model reactions (Eqs. 1,2,5) can be described as follows:

31 2 4

2 2 2

5

6 7 8 9

2 2 2

mm m 3mSRM R MeOH H O R CO H

mMD D MeOH

m m m mRWGS W CO H W CO H O

r k p p k p p (12

r k p (13

r k p p k p p (14)

−

−

= ⋅ ⋅ − ⋅ ⋅

= ⋅

= ⋅ ⋅ − ⋅ ⋅

)

)

Accordingly, the differential equations for the single

components (obtained by the Maple V software, release 4.00c) are given below.



8

2 2

2 2

2

MeOHSRM MD

H O H OSRM RWGS SRM MD

MeOH,0

CO COMD RWGS SRM MD

MeOH,0

CO COSRM RWGS SRM MD

MeOH,0

HSRM MD RWGS SRM

p 2(r r ) (15)t 1 X

p p(r r ) (r r ) (16)

t pp p(r r ) (r r ) (17)

t pp p

(r r ) (r r ) (18)t p

p(3r 2r r ) (r

t

∂= − + ⋅

∂ +∂

= − + − + ⋅∂

∂= + − + ⋅

∂

∂= − − + ⋅

∂

∂= + − − +

∂2H

MDMeOH,0

pr ) (19)

p⋅

f

f

ln(1 X ) (23)−ω = −

l

The kinetic evaluation was performed by means of

the Berkeley Madonna 8.0.1. software. A Runge Kutta method was used to solve the differential equations (Eqs. 15-19) and the experimental data were fitted by a least-squares method. The rate constants for the model reactions were determined by fitting the simulation to the experimen-tal data by varying the reaction orders until a good agree-ment was obtained and this procedure was repeated until an optimal fit was achieved. According to the previous results of Purnama et al., the total reaction order for SRM (m1+m2) was 1, whereas the individual reaction orders determined experimentally for methanol and water were 0.6 and 0.4, respectively [21]. The application of the above reaction orders ensured an optimal fit for the current experimental data. For the components involved in the RWGS reaction, the reaction orders were set to 1 [21], in correlation with those published by Choi and Stenger in their detailed study on the kinetics of the WGS reaction [48], including rate expressions derived from various reaction mechanisms [49]. The rate constants for both the RWGS and the MD reactions were considerably smaller than that of SRM, as revealed by the low CO production, and thus the reaction orders of the components implied in RWGS and MD had no appreciable effect on the fitting results. The reaction orders are summarized in Table 2.

The increased mole numbers during SRM (4 mole of

products formed from 2 mole of reactants) were taken into account via expressing the conversion X by the partial pres-sures of methanol as

MeOH,0 MeOH,t

MeOH,0 MeOH,t

p pX (20)

p p−

=+

Likewise, the contact time τ was determined by con-

sidering the increase of volume in the catalyst bed during conversion. It was calculated by integrating the progress of

conversion as a function of the reactor length l

inside the catalyst bed, as described in Eq. 21,

X( )l

00 0

1 A 1( )d d (21)w 1 X( )

τ = τ =+∫ ∫

l l

l l ll l

The partial pressures for the components of the prod-uct stream were determined as a function of the contact time at the temperatures 503, 523, 543 and 573 K for all catalysts. The results obtained for CZC15 are displayed in Figs.7-10. Between the respective data for the other CZC samples, no significant differences were observed.

where A is the cross section of the reactor and is the

initial flow rate of the reactants. Further, may be

sufficiently well approximated by an exponential function (Eq.22),

0wX( )l

It can be seen from Figs. 7-10 that the partial pres-sures of the reactants methanol and water decreased to the same extent with increasing contact times at all tempera-tures. The most pronounced decrease can be observed at 573 K, where complete transformation occurred at τ < 1 s.

X( ) 1 e (22)−ω= − ll

and the parameter ω can be obtained from the values of the final conversion Xf and the final length of the catalyst bed

(Eq.23). fl

Table 2: Reaction orders for the components involved in the kinetic model

SRM RWGS MD

component reaction order component reaction order component reaction order

MeOH 0,6 CO2 1,0 MeOH 1,3

H2O 0,4 H2 1,0

CO2 1,0 CO 1,0

H2 1,0 H2O 1,0



9

0

20

40

60

80

0 1 2 3 4 5 6

τ [s]

p [k

Pa]

0

100

200

300

400

p CO [P

a]

H2CO2H2OMeOHCO

Figure 7: Partial pressures of the components in the reac-tion mixture during SRM on CZC15, experimental data and fitting results m = 0.6726 g, T = 503 K, p = 105 Pa, Me-OH:H2O = 1:1

0

20

40

60

80

0 0.5 1 1.5 2 2.5

τ [s]

p [k

Pa]

0

150

300

450

600

p CO [P

a]

H2CO2H2OMeOHCO


0

20

40

60

80

0 0.2 0.4 0.6 0.8 1

τ [s]

p [k

Pa]

0

100

200

300

400

p CO [P

a]

H2CO2H2OMeOHCO


0

20

40

60

80

0 0.2 0.4 0.6 0.8 1

τ [s]

p [k

Pa]

0

400

800

1200

1600

p CO [P

a]

H2CO2H2OMeOHCO


Meanwhile, the partial pressures of the main reaction products H2 and CO2 increased steadily during reaction, and could be fitted with saturation curves, for which the limits for H2 and CO2 production proved to be 70-75 kPa and 24 kPa, respectively. This implies that the amount of H2 gained in the reaction at 573 K attains the theoretical maximum of 75%, corresponding to the thermodynamical equilibrium state [4], and approaches it reasonably well at lower temperatures. The differences in the partial pressures of H2 and CO2 at various contact times confirm that the amount of H2 formed in the reaction was 3 times higher than that of CO2, irrespective of the reaction temperature and the conversion level. By increasing the contact time, CO2 formation was less affected than H2 production and the contact time required to reach the saturation value de-creased with the reaction temperature to an appreciable extent.

Compared to a commercial Cu/ZnO/Al2O3 catalyst [21], the H2 partial pressures obtained for the CZC samples were somewhat lower (74–85% of those reported for the commercial sample between 503 K and 573 K, respec-tively) at short contact times. Nevertheless, the saturation values observed for the CZC catalysts at long contact times approached 75 kPa at all temperatures.

The minor amount of CO detected in the reaction mixture was found to increase with both the contact time and the reaction temperature. In the temperature range in-vestigated, the amount of CO typically varied between 0.06-0.4%, only exceeding that at 573 K, at contact times longer than 0.4 s. Unlike the main reaction products, the formation of CO can be described with S-shaped curves at each temperature, which may be regarded as a further evi-dence of CO being a secondary product in the reaction [6,21]. As shown in Figs 7-10, the partial pressures of CO could be fitted at all temperatures and contact times when MD was included in the reaction scheme. Nevertheless, it should be noted that the significance of MD can be ob-served merely at low contact times and therefore CO pro-duction is suggested to occur predominantly by the RWGS



10 reaction. The conversions obtained for CZC15 at different temperatures, plotted as a function of the contact time, are displayed in Fig. 11.

0

0.2

0.4

0.6

0.8

1

0 2 4 6 8 10

τ [s]

Con

vers

ion

503 K523 K543 K573 K

Figure 11: Methanol conversions during SRM of CZC15 as a function of contact time at different reaction tempera-tures (m = 0.6726 g, p = 105 Pa, MeOH:H2O = 1:1)

The complete transformation of methanol could be achieved at all reaction temperatures, except for 503 K, at which the contact time corresponding to full conversion was determined by extrapolation (τ ∼10 s). It can be seen that an increase in the reaction temperature resulted in a marked decrease of the contact time corresponding to the same conversion level. Accordingly, the full conversion of methanol at 573 K was observed at a contact time of 0.92 s. Although an increase in the reaction temperature proved to be beneficial in terms of the catalytic activity, as suggested by Fig.11, it also resulted in the formation of an increased amount of CO in the product mixture (see Fig.12). Fur-thermore, a prolonged exposure of the catalyst to 573 K was found to result in a moderate deactivation, which may be due to sintering of the active Cu particles [6,47]. For the other CZC samples, similar observations have been made. Likewise, Lindström and Pettersson reported that sintering of Cu particles of a Cu/ZnO/Al2O3 catalyst subjected to 613 K resulted in a loss of Cu surface area [50]. Hence, the optimal reaction temperature range for the steam reforming of methanol on CZC catalysts is 523-543 K, at which high methanol conversions can be achieved while maintaining a low CO level.

For the CO partial pressures, fitting of the experi-mental data was performed by setting the final values, cal-culated at long contact times from the thermodynamical equilibrium constant of the RWGS reaction, as displayed in Fig.12. For comparison, these values were also determined experimentally by investigating a large amount of catalyst (m = 3.55 g) at a low flow rate of 0.01 cm3min-1 to ensure an extended contact time for the reaction. The CO partial pressures obtained from repeated runs at different tempera-tures were in good agreement with the theoretical values. It can be observed in Fig. 12 that the experimental CO levels are located at the low end of the theoretical curves and, hence, are far below those predicted by equilibrium calcu

503 K

3.5 kPa

523 K

4.2 kPa543 K

4.8 kPa

573 K5.9 kPa

0

2

4

6

0 30 60 90 120 150

τ [s]

p CO [k

Pa]

503 K523 K543 K573 K

Figure 12: CO partial pressures during SRM of CZC15 as a function of contact time at different reaction temperatures (m = 0.6726 g, p = 105 Pa, MeOH:H2O = 1:1)

lations at all temperatures. It follows that the limiting CO partial pressures could be attained only at contact times considerably longer than those obtained under the present experimental conditions (see also Figs. 7-10). Accordingly, it is reasonable to assume that the kinetics of the CO forma-tion is more complex than that described within the frame-work of this study [21]. Nevertheless, the kinetic model employed provides a good approximation for the experi-mental data obtained for all components in the contact time region where complete methanol conversion occurs, which may be of major interest for industrial applications.

The reaction rate constants k for the individual reac-tions, obtained from the simulations performed by applying the reaction orders indicated in Table 2, are listed below in Table 3. For a better comparison, all values were referred to a CuO content of 1.0 g.

For the most active catalyst (CZC35) investigated at the highest reaction temperature (573 K), the Thiele modulus [51] was calculated to be 0.594, from which the mass transfer factor (the actual reaction rate as referred to the ideal reaction rate without transport limitation) was 0.977, indicating that the mass transport limitation was indeed negligible and thus the reaction was kinetically con-trolled. Given that the reaction rates for the other CZC samples were found to be typically lower, in particular at temperatures below 573 K, the effect of intraparticle trans-port phenomena on the kinetic parameters can be safely excluded.

The apparent activation energies were determined from the slopes of the Arrhenius plots, as expressed in Eq. 24.

A

0Eln k ln k (24)RT

= −

A typical illustration of the Arrhenius parameters for

the model reactions on CZC15 is displayed in Fig.13. It may be readily observed from the slopes that the activation



11 Table 3: Reaction rate constants determined for the SRM, RWGS and MD reactions on the CZC catalysts at different reaction

temperatures

Sample kR [s-1g-1]a∗ k-W [bar-1s-1g-1]b∗ kD [bar-0.3s-1g-1]c∗

503K 523K 543K 573K 503K 523K 543K 573 K 503K 523K 543K 573K

CZC5 2.92 10.71 34.40 68.44 0.04 0.11 0.30 1.12 0.04 0.13 0.66 2.12

CZC15 22.58 50.67 117.33 268.02 0.27 0.60 1.28 3.55 0.09 0.36 1.02 2.89

CZC25 20.25 55.17 79.95 146.37 0.22 0.95 1.35 3.06 0.14 0.38 1.13 4.50

CZC35 26.67 61.33 89.00 165.69 0.31 0.58 1.54 2.31 0.36 0.96 1.54 3.44

a reaction rate constant of SRM b reaction rate constant of RWGS c reaction rate constant of MD ∗ all values were referred to a CuO content of 1.0 g Table 4: Apparent activation energies for the SRM, RWGS and MD reactions on the CZC catalysts

Reaction EA [kJmol-1]

CZC5

EA [kJmol-1]

CZC15

EA [kJmol-1]

CZC25

EA [kJmol-1]

CZC35

SRM 109.2 85.4 66.1 60.7

RWGS 115.5 88.2 85.7 69.3

MD 142.6 118.7 104.9 75.5

energies of SRM and RWGS are comparable and consid-erably lower than that of MD. The values of the activation energies for the SRM, RWGS and MD reactions, deter-mined for all the CZC samples, are listed in Table 4.

According to Table 4, no significant differences be-tween the activation energies of the SRM and RWGS reac-tions can be observed, except for CZC25. For the latter sample, the activation energy of the RWGS reaction was considerably higher than that for SRM, which may account for the particularly low CO level observed for this catalyst (see Fig.4). On the other hand, the activation energies for MD proved to be considerably higher than those of SRM and RWGS and therefore MD may be regarded as the rate-limiting step for CO formation at short contact times. As mentioned above, the significance of MD in the overall reaction mechanism decreases dramatically at contact times exceeding 0.3-1 s, depending on the reaction temperature. Accordingly, the enhanced CO levels obtained at longer contact times can be attributed to the predominance of the RWGS reaction, which requires a considerably lower acti-vation energy.

For the main reaction (SRM), the apparent activation energies proved to be relatively close for all samples, apart from CZC5, for which the highest value was obtained. For comparison, the Arrhenius plots for all the CZC samples are depicted in Fig.14. For CZC5 and CZC15, the apparent

activation energies are in good agreement with those re-ported in the literature for Cu/ZnO/Al2O3 catalysts, whereas the values for CZC25 and CZC35 are lower [6,16,21,52].

0

1

2

3

4

5

6

1.72 1.82 1.92 2.02

10-3T-1 [K-1]

ln k

SRM

CZC 5CZC 15CZC 25CZC 35

Figure 14: Arrhenius representation of the reaction rate constants obtained for the SRM reaction on the CZC cata-lysts

For all the model reactions (SRM, RWGS, and MD), the activation energies determined exhibit a linear correla-tion with the Cu content, as shown in Fig.15. On increasing the amount of Cu, a systematic decrease can be observed for each reaction. The most pronounced decrease can be observed in the activation energy of MD, indicating that at



12 lower Cu loading the contribution of MD to CO formation diminishes. Takeguchi et al. have reported a similar trend for the SRM reaction on two Cu/ZnO catalysts [53], al-though a potential mass transport limitation was not taken into account and thus their studies cannot be directly com

40

80

120

160

0 10 20 30

Cu content [%]

E A [k

Jmol

-1]

40

SRMRWGSMD

Figure 15: Apparent activation energies of the model reac-tions (SRM, RWGS, and MD) as a function of the Cu con-tent of four CZC catalysts pared with the results presented here. Conversely, in the current study, the well-known effect of mass transport limi-tation on the kinetic studies to result in a reduced apparent activation energy can be excluded. Therefore, it is con-cluded that the variation of the Cu concentration in the CeO2/ZrO2 precursor material gave rise to the formation of CZC samples with significantly different catalytic behav-iour. Whereas the chemical complexity in the various CZC catalysts remains the same, the chemical composition ad-justed in the preparation procedure results in a significantly altered structural complexity of the materials. These cata-lysts with vastly different active surfaces exhibit a catalytic behaviour that is not simply correlated to the accessible surface area. Dispersion, Cu particle size, and Cu support interactions have a pronounced influence on the micro-structure of the Cu particles and the effect of the micro-structure on the active Cu surface. Eventually, both the Cu surface area and the microstructural properties of the Cu

particles need to be elucidated in order to fully understand the significantly different catalytic activities and the kinet-ics described above. 4.Conclusions

Novel Cu/ZrO2/CeO2 materials prepared by co-precipitation and polymer templating have been investi-gated as catalysts in the steam reforming of methanol. Cata-lytic measurements were performed under continuous operation in a fixed bed reactor at atmospheric pressure. Time-on-stream experiments indicated that following an initial period of deactivation, the conversions for the sam-ples with Cu contents exceeding 5% were stabilized and thus these samples could be suitable for prolonged catalytic applications. An increase in the Cu content from 5 to 15% was found to improve the long-term stability and suppress the CO production considerably, whereas the effect of a further enhancement was less significant. With respect to a reduced formation of CO during SRM, the samples with increased Cu loadings proved to be more efficient catalysts.

Since the effect of mass transport limitations on the catalytic reaction was found to be insignificant, a kinetic analysis was undertaken. The reaction scheme suggested for the overall transformation included the SRM, MD and RWGS reactions. The experimental data obtained in the range 503-573 K could be well fitted by the kinetic model employed. The optimal temperature range for SRM on the CZC catalysts was 523-543 K, for which high methanol conversions and low CO levels were obtained. For the SRM and RWGS reactions, the values of the activation energies were comparable and considerably lower than that of MD. The activation energy of each individual reaction exhibited a systematic decrease with an increasing Cu con-tent of the catalyst. Evidently, the variation in the Cu con-centration of the precursor material during preparation altered the microstructure of the Cu particles and, thus the active Cu surface, which considerably affected the catalytic behaviour of the CZC catalysts.

References [1] B. Lindström, L.J. Pettersson, Int. J. Hydrogen Energy 26

(2001) 923. [2] S. Velu, K. Suzuki, M. Okazaki, M.P. Kapoor, T. Osaki, F.

Ohashi, J. Catal. 194 (2000) 373. [3] J.P. Breen, J.R.H. Ross, Catal. Today 51 (1999) 521. [4] P.J. de Wild, M.J.F.M. Verhaak, Catal. Today 60 (2000) 3. [5] J. Agrell, H. Birgersson, M. Boutonnet, I. Melian-Cabrera,

R.M. Navarro, J.L.G. Fierro, J. Catal. 219 (2003) 389. [6] J. Agrell, H. Birgersson, M. Boutonnet, J. Power Sources,

4654 (2002) 1. [7] G.A. Olah, Catal. Lett. 93 (2004) 1. [8] B. Lindström, L.J. Pettersson, J. Power Sources 106 (2002)

264. [9] J. Agrell, M. Boutonnet, J.L.G. Fierro, Appl. Catal. A 253

(2003) 213. [10] S. Velu, K. Suzuki, T. Osaki, Chem. Commun. (1999) 2341.

[11] J. Papavasiliou, G. Avgouropoulos, T. Ioannides, Catal. Commun. 5 (2004) 231.

[12] X.R. Zhang, P. Shi, J. Zhao, M. Zhao, C. Liu, Fuel Process-ing Technol. 83 (2003) 183.

[13] [13] K. Takahashi, N. Takezawa, H. Kobayashi, Appl. Catal. 2 (1982) 383.

[14] E. Santacesaria, S. Carra, Appl. Catal. 5 (1983) 345. [15] J.C. Amphlett, R.F. Mann, R.D. Weir, Can. J. Chem. Eng. 66

(1988) 950. [16] C.J. Jiang, D.L. Trimm, M.S. Wainwright, Appl. Catal. A 93

(1993) 245. [17] N. Iwasa, S. Masuda, N. Ogawa, N. Takezawa, Appl. Catal.

A 125 (1995) 145. [18] C. Cao, G. Xia, J. Holladay, E. Jones, Y. Wang, Appl. Catal.

A 262 (2004) 19.



13 [19] W.H. Cheng, I. Chen, J. Liou, S.S. Lin, Top. Catal. 22 (2003)

225. [20] M.V. Twigg, M.S. Spencer, Top. Catal. 22 (2003) 191. [21] H. Purnama, T. Ressler, R.E. Jentoft, H. Soerijanto, R.

Schlögl, R. Schomäcker, Appl. Catal. A 259 (2004) 83. [22] M.M. Günter, T. Ressler, R.E. Jentoft, B. Bems, J. Catal. 203

(2001) 133. [23] B.L. Kniep, T. Ressler, A. Rabis, F. Girgsdies, M. Baenitz, F.

Steglich, R. Schlögl, Angew. Chem. Int. Ed. 43 (2004) 112. [24] T. Shishido, Y. Yamamoto, H. Morioka, K. Takaki, K. Take-

hira, Appl. Catal. A 263 (2004) 249. [25] B.A. Peppley, J.C. Amphlett, L.M. Kearns, R.F. Mann, Appl.

Catal. A 179 (1999) 21. [26] S.P. Asprey, B.W. Wojciechowski, B.A. Peppley, Appl.

Catal. A 179 (1999) 51. [27] H. Purnama, F. Girgsdies, T. Ressler, J.H. Schattka, R.A.

Caruso, R. Schomäcker, R. Schlögl, Catal. Lett. 94 (2004) 61.

[28] J.P. Breen, F.C. Meunier, R.H. Ross, Chem. Commun. (1999) 2247.

[29] P.H. Matter, D.J. Braden, U.S. Ozkan, J. Catal. 223 (340) 2004.

[30] X. Zhang, P. Shi, J. Mol. Catal. A 194 (2003) 99. [31] Y. Liu, T. Hayakawa, K. Suzuki, S. Hamakawa, T. Tsunoda,

T. Ishii, M. Kumagai, Appl. Catal. A 223 (2002) 137. [32] H.S. Roh, K.W. Jun, W.S. Dong, S.E. Park, Y.S. Baek, Catal.

Lett. 74 (2001) 31. [33] M.F. Luo, J. Chen, L.S. Chen, J.Q. Lu, Z. Feng, C. Li, Chem.

Mater. 13 (2001) 197. [34] B.C. Gates, Catalytic Chemistry, Wiley, New York, 1992. [35] A.S. Deshpande, N. Pinna, P. Beato, M. Antonietti, M. Nie-

derberger, Chem. Mater. 16 (2004) 2599.

[36] G.C. Chinchen, C.M. Hay, H.D. Vandervell, K.C. Waugh, J. Catal. 103 (1987) 79.

[37] M.M. Günter, T. Ressler, B. Bems, C. Büscher, T. Genger, O. Hinrichsen, M. Muhler, R. Schlögl, Catal. Lett. 71 (2001) 37.

[38] R.A. Young, The Rietveld Method, Oxford University Press, New York, 1993.

[39] Y. Choi, H.G. Stenger, Appl. Catal. B 38 (2002) 259. [40] J. Agrell, M. Boutonnet, I. Melian-Cabrera, J.L.G. Fierro,

Appl. Catal. A 253 (2003) 201. [41] E.D. Schrum, T.L. Reitz, H.H. Kung, Stud. Surf. Sci. Catal.

139 (2001) 229. [42] J.C. Amphlett, M.J. Evans, R.F. Mann, R.D. Weir, Can. J.

Chem. Eng. 63 (1985) 605. [43] J.C. Amphlett, R.F. Mann, B.A. Peppley, Stud. Surf. Sci.

Catal. 81 (1994) 409. [44] H. Kobayashi, N. Takezawa, C. Minochi, Chem. Lett. (1976)

1347. [45] K. Takahashi, N. Takezawa, H. Kobayashi, Appl. Catal. 2

(1982) 383. [46] M.S. Spencer, Nature 323 (1986) 685. [47] M.V. Twigg, M.S. Spencer, Appl. Catal. A 212 (2001) 161. [48] Y. Choi, H.G. Stenger, J. Power Sources 124 (2003) 432. [49] N. Amadeo, M. Laborde, Int. J. Hydrogen Energy 20 (1995)

949. [50] B. Lindström, L.J. Pettersson, Catal. Lett. 74 (2001) 27. [51] O. Levenspiel, Chemical Reaction Engineering, Wiley, New

York, 1972. [52] K. Geissler, E. Newson, F. Vogel, T.B. Truong, P. Hottinger,

A. Wokaun, Phys. Chem. Chem. Phys. 3 (2001) 47. [53] T. Takeguchi, Y. Kani, M. Inoue, K. Eguchi, Catal. Lett. 83

(2002) 49.

Documents

Steam reforming of methanol over Cu/ZrO2/CeO2 …pubman.mpdl.mpg.de/pubman/item/escidoc:738831:1/component/escido… · Steam reforming of methanol over Cu/ZrO2/CeO2 catalysts: a